## The No Nonsense Guide to Getting Cited by AI *By Ethan Young | Updated July 1, 2025* %%[[The No BS Guide to Getting Cited by AI 1-pager at 56 downscale percent v3]]%% ### Executive Summary **Not all AI citations come from page 1, let alone rank 1**—a shift that calls for new brand visibility strategies. Brands that consistently create and get mentioned in authoritative content will maintain visibility as AI search evolves. ### **What If Ranking #5 Could Outperform #1 in AI Overview?** While highly correlated, AI Overviews demonstrates a surprising disconnect from traditional organic search rankings. [Petrescu (2024)](https://www.advancedwebranking.com/blog/ai-overview-study) analyzed 8,000 keywords across 16 industries and found: - **33.4%** of AI Overview links rank in the top 10 organic results - **46.5%** of cited URLs rank outside the top 50 organic results ### **What Works: Generative Engine Optimization (GEO)** [Aggarwal et al. (2024)](https://arxiv.org/abs/2311.09735) analyzed 10,000 queries through a simulated generative engine—defined as an LLM + external database—built with GPT-3.5-turbo and found 6 methods that increased AI visibility by 15-40%, with real-world validation on a subset of 200 queries tested through Perplexity.ai showing improvements of up to 37%: 1. **Cite Sources** → Include citations from reliable sources 2. **Add Statistics** → Incorporate relevant data points 3. **Include Quotes** → Add credible expert commentary 4. **Optimize Fluency** → Ensure high-quality writing 5. **Simplify Language** → Make content easy to understand 6. **Avoid Keyword Stuffing** → Don't over-optimize **The best combination** (Optimize Fluency & Add Statistics) outperformed single methods by 5.5%, though domain-specific optimization is possible. #### **GEO Strategy by Ranking Position** Aggarwal et al. (2024) found dramatic visibility shifts when lower-ranked competitors used GEO's **Cite Sources** method: - **Challenger Brands:** Aggressively adopt GEO (5th-ranked sites gained 115% visibility) - **Established Brands:** Defend with GEO adoption (top sites lost 30% without it) ### **Why Existing Frameworks Fall Short** Most frameworks treat AI as a monolithic technology when it's actually 3 distinct architectures that retrieve information differently: 1. **Standalone LLM** → trained on datasets with knowledge cutoff dates (Anthropic's Claude 3.5 Sonnet, OpenAI's GPT-4) 2. **LLM-first + search** → LLM calls search engine when needed (Claude 4 Sonnet, GPT-4o w/ browsing, Perplexity AI) 3. **Search-first + LLM** → search engine results recapped by LLM (Google’s AI Overview/AI Mode w/ Gemini, Bing Chat w/ OpenAI) ### **The 3 Pillars of AI Representation** To improve AI representation across standalone LLMs and generative engines: 1. **Master SEO fundamentals** (then adapt for crawlers like OAI-SearchBot, ClaudeBot, & PerplexityBot that don't execute JavaScript) 2. **Create and retrofit existing high-performing content** with GEO methods (prioritize decision-making & proprietary data) 3. **Earn authoritative mentions** (prioritize Wikipedia and major news outlets) that simultaneously: - **Shape LLM training data** over time (future AI models) - **Boost search rankings** through backlinks (immediate SEO benefit) - **Provide credible sources** for generative engines **Have questions/concerns?** [email protected]